#include <linux/ptrace.h>
#include <linux/audit.h>
#include <linux/stddef.h>
+#include <linux/slab.h>
+#include <linux/security.h>
#include <asm/uaccess.h>
#include <asm/io.h>
#include <asm/tlbflush.h>
#include <asm/irq.h>
+#include <asm/traps.h>
+#include <asm/vm86.h>
/*
* Known problems:
*/
-#define KVM86 ((struct kernel_vm86_struct *)regs)
-#define VMPI KVM86->vm86plus
-
-
/*
* 8- and 16-bit register defines..
*/
/*
* virtual flags (16 and 32-bit versions)
*/
-#define VFLAGS (*(unsigned short *)&(current->thread.v86flags))
-#define VEFLAGS (current->thread.v86flags)
+#define VFLAGS (*(unsigned short *)&(current->thread.vm86->veflags))
+#define VEFLAGS (current->thread.vm86->veflags)
#define set_flags(X, new, mask) \
((X) = ((X) & ~(mask)) | ((new) & (mask)))
#define SAFE_MASK (0xDD5)
#define RETURN_MASK (0xDFF)
-/* convert kernel_vm86_regs to vm86_regs */
-static int copy_vm86_regs_to_user(struct vm86_regs __user *user,
- const struct kernel_vm86_regs *regs)
-{
- int ret = 0;
-
- /*
- * kernel_vm86_regs is missing gs, so copy everything up to
- * (but not including) orig_eax, and then rest including orig_eax.
- */
- ret += copy_to_user(user, regs, offsetof(struct kernel_vm86_regs, pt.orig_ax));
- ret += copy_to_user(&user->orig_eax, ®s->pt.orig_ax,
- sizeof(struct kernel_vm86_regs) -
- offsetof(struct kernel_vm86_regs, pt.orig_ax));
-
- return ret;
-}
-
-/* convert vm86_regs to kernel_vm86_regs */
-static int copy_vm86_regs_from_user(struct kernel_vm86_regs *regs,
- const struct vm86_regs __user *user,
- unsigned extra)
-{
- int ret = 0;
-
- /* copy ax-fs inclusive */
- ret += copy_from_user(regs, user, offsetof(struct kernel_vm86_regs, pt.orig_ax));
- /* copy orig_ax-__gsh+extra */
- ret += copy_from_user(®s->pt.orig_ax, &user->orig_eax,
- sizeof(struct kernel_vm86_regs) -
- offsetof(struct kernel_vm86_regs, pt.orig_ax) +
- extra);
- return ret;
-}
-
-struct pt_regs *save_v86_state(struct kernel_vm86_regs *regs)
+void save_v86_state(struct kernel_vm86_regs *regs, int retval)
{
struct tss_struct *tss;
- struct pt_regs *ret;
- unsigned long tmp;
+ struct task_struct *tsk = current;
+ struct vm86plus_struct __user *user;
+ struct vm86 *vm86 = current->thread.vm86;
+ long err = 0;
/*
* This gets called from entry.S with interrupts disabled, but
*/
local_irq_enable();
- if (!current->thread.vm86_info) {
- pr_alert("no vm86_info: BAD\n");
+ if (!vm86 || !vm86->user_vm86) {
+ pr_alert("no user_vm86: BAD\n");
+ do_exit(SIGSEGV);
+ }
+ set_flags(regs->pt.flags, VEFLAGS, X86_EFLAGS_VIF | vm86->veflags_mask);
+ user = vm86->user_vm86;
+
+ if (!access_ok(VERIFY_WRITE, user, vm86->vm86plus.is_vm86pus ?
+ sizeof(struct vm86plus_struct) :
+ sizeof(struct vm86_struct))) {
+ pr_alert("could not access userspace vm86 info\n");
do_exit(SIGSEGV);
}
- set_flags(regs->pt.flags, VEFLAGS, X86_EFLAGS_VIF | current->thread.v86mask);
- tmp = copy_vm86_regs_to_user(¤t->thread.vm86_info->regs, regs);
- tmp += put_user(current->thread.screen_bitmap, ¤t->thread.vm86_info->screen_bitmap);
- if (tmp) {
- pr_alert("could not access userspace vm86_info\n");
+
+ put_user_try {
+ put_user_ex(regs->pt.bx, &user->regs.ebx);
+ put_user_ex(regs->pt.cx, &user->regs.ecx);
+ put_user_ex(regs->pt.dx, &user->regs.edx);
+ put_user_ex(regs->pt.si, &user->regs.esi);
+ put_user_ex(regs->pt.di, &user->regs.edi);
+ put_user_ex(regs->pt.bp, &user->regs.ebp);
+ put_user_ex(regs->pt.ax, &user->regs.eax);
+ put_user_ex(regs->pt.ip, &user->regs.eip);
+ put_user_ex(regs->pt.cs, &user->regs.cs);
+ put_user_ex(regs->pt.flags, &user->regs.eflags);
+ put_user_ex(regs->pt.sp, &user->regs.esp);
+ put_user_ex(regs->pt.ss, &user->regs.ss);
+ put_user_ex(regs->es, &user->regs.es);
+ put_user_ex(regs->ds, &user->regs.ds);
+ put_user_ex(regs->fs, &user->regs.fs);
+ put_user_ex(regs->gs, &user->regs.gs);
+
+ put_user_ex(vm86->screen_bitmap, &user->screen_bitmap);
+ } put_user_catch(err);
+ if (err) {
+ pr_alert("could not access userspace vm86 info\n");
do_exit(SIGSEGV);
}
tss = &per_cpu(cpu_tss, get_cpu());
- current->thread.sp0 = current->thread.saved_sp0;
- current->thread.sysenter_cs = __KERNEL_CS;
- load_sp0(tss, ¤t->thread);
- current->thread.saved_sp0 = 0;
+ tsk->thread.sp0 = vm86->saved_sp0;
+ tsk->thread.sysenter_cs = __KERNEL_CS;
+ load_sp0(tss, &tsk->thread);
+ vm86->saved_sp0 = 0;
put_cpu();
- ret = KVM86->regs32;
+ memcpy(®s->pt, &vm86->regs32, sizeof(struct pt_regs));
- ret->fs = current->thread.saved_fs;
- set_user_gs(ret, current->thread.saved_gs);
+ lazy_load_gs(vm86->regs32.gs);
- return ret;
+ regs->pt.ax = retval;
}
static void mark_screen_rdonly(struct mm_struct *mm)
static int do_vm86_irq_handling(int subfunction, int irqnumber);
-static void do_sys_vm86(struct kernel_vm86_struct *info, struct task_struct *tsk);
+static long do_sys_vm86(struct vm86plus_struct __user *user_vm86, bool plus);
-SYSCALL_DEFINE1(vm86old, struct vm86_struct __user *, v86)
+SYSCALL_DEFINE1(vm86old, struct vm86_struct __user *, user_vm86)
{
- struct kernel_vm86_struct info; /* declare this _on top_,
- * this avoids wasting of stack space.
- * This remains on the stack until we
- * return to 32 bit user space.
- */
- struct task_struct *tsk = current;
- int tmp;
-
- if (tsk->thread.saved_sp0)
- return -EPERM;
- tmp = copy_vm86_regs_from_user(&info.regs, &v86->regs,
- offsetof(struct kernel_vm86_struct, vm86plus) -
- sizeof(info.regs));
- if (tmp)
- return -EFAULT;
- memset(&info.vm86plus, 0, (int)&info.regs32 - (int)&info.vm86plus);
- info.regs32 = current_pt_regs();
- tsk->thread.vm86_info = v86;
- do_sys_vm86(&info, tsk);
- return 0; /* we never return here */
+ return do_sys_vm86((struct vm86plus_struct __user *) user_vm86, false);
}
SYSCALL_DEFINE2(vm86, unsigned long, cmd, unsigned long, arg)
{
- struct kernel_vm86_struct info; /* declare this _on top_,
- * this avoids wasting of stack space.
- * This remains on the stack until we
- * return to 32 bit user space.
- */
- struct task_struct *tsk;
- int tmp;
- struct vm86plus_struct __user *v86;
-
- tsk = current;
switch (cmd) {
case VM86_REQUEST_IRQ:
case VM86_FREE_IRQ:
}
/* we come here only for functions VM86_ENTER, VM86_ENTER_NO_BYPASS */
- if (tsk->thread.saved_sp0)
- return -EPERM;
- v86 = (struct vm86plus_struct __user *)arg;
- tmp = copy_vm86_regs_from_user(&info.regs, &v86->regs,
- offsetof(struct kernel_vm86_struct, regs32) -
- sizeof(info.regs));
- if (tmp)
- return -EFAULT;
- info.regs32 = current_pt_regs();
- info.vm86plus.is_vm86pus = 1;
- tsk->thread.vm86_info = (struct vm86_struct __user *)v86;
- do_sys_vm86(&info, tsk);
- return 0; /* we never return here */
+ return do_sys_vm86((struct vm86plus_struct __user *) arg, true);
}
-static void do_sys_vm86(struct kernel_vm86_struct *info, struct task_struct *tsk)
+static long do_sys_vm86(struct vm86plus_struct __user *user_vm86, bool plus)
{
struct tss_struct *tss;
-/*
- * make sure the vm86() system call doesn't try to do anything silly
- */
- info->regs.pt.ds = 0;
- info->regs.pt.es = 0;
- info->regs.pt.fs = 0;
-#ifndef CONFIG_X86_32_LAZY_GS
- info->regs.pt.gs = 0;
-#endif
+ struct task_struct *tsk = current;
+ struct vm86 *vm86 = tsk->thread.vm86;
+ struct kernel_vm86_regs vm86regs;
+ struct pt_regs *regs = current_pt_regs();
+ unsigned long err = 0;
+
+ err = security_mmap_addr(0);
+ if (err) {
+ /*
+ * vm86 cannot virtualize the address space, so vm86 users
+ * need to manage the low 1MB themselves using mmap. Given
+ * that BIOS places important data in the first page, vm86
+ * is essentially useless if mmap_min_addr != 0. DOSEMU,
+ * for example, won't even bother trying to use vm86 if it
+ * can't map a page at virtual address 0.
+ *
+ * To reduce the available kernel attack surface, simply
+ * disallow vm86(old) for users who cannot mmap at va 0.
+ *
+ * The implementation of security_mmap_addr will allow
+ * suitably privileged users to map va 0 even if
+ * vm.mmap_min_addr is set above 0, and we want this
+ * behavior for vm86 as well, as it ensures that legacy
+ * tools like vbetool will not fail just because of
+ * vm.mmap_min_addr.
+ */
+ pr_info_once("Denied a call to vm86(old) from %s[%d] (uid: %d). Set the vm.mmap_min_addr sysctl to 0 and/or adjust LSM mmap_min_addr policy to enable vm86 if you are using a vm86-based DOS emulator.\n",
+ current->comm, task_pid_nr(current),
+ from_kuid_munged(&init_user_ns, current_uid()));
+ return -EPERM;
+ }
+
+ if (!vm86) {
+ if (!(vm86 = kzalloc(sizeof(*vm86), GFP_KERNEL)))
+ return -ENOMEM;
+ tsk->thread.vm86 = vm86;
+ }
+ if (vm86->saved_sp0)
+ return -EPERM;
+
+ if (!access_ok(VERIFY_READ, user_vm86, plus ?
+ sizeof(struct vm86_struct) :
+ sizeof(struct vm86plus_struct)))
+ return -EFAULT;
+
+ memset(&vm86regs, 0, sizeof(vm86regs));
+ get_user_try {
+ unsigned short seg;
+ get_user_ex(vm86regs.pt.bx, &user_vm86->regs.ebx);
+ get_user_ex(vm86regs.pt.cx, &user_vm86->regs.ecx);
+ get_user_ex(vm86regs.pt.dx, &user_vm86->regs.edx);
+ get_user_ex(vm86regs.pt.si, &user_vm86->regs.esi);
+ get_user_ex(vm86regs.pt.di, &user_vm86->regs.edi);
+ get_user_ex(vm86regs.pt.bp, &user_vm86->regs.ebp);
+ get_user_ex(vm86regs.pt.ax, &user_vm86->regs.eax);
+ get_user_ex(vm86regs.pt.ip, &user_vm86->regs.eip);
+ get_user_ex(seg, &user_vm86->regs.cs);
+ vm86regs.pt.cs = seg;
+ get_user_ex(vm86regs.pt.flags, &user_vm86->regs.eflags);
+ get_user_ex(vm86regs.pt.sp, &user_vm86->regs.esp);
+ get_user_ex(seg, &user_vm86->regs.ss);
+ vm86regs.pt.ss = seg;
+ get_user_ex(vm86regs.es, &user_vm86->regs.es);
+ get_user_ex(vm86regs.ds, &user_vm86->regs.ds);
+ get_user_ex(vm86regs.fs, &user_vm86->regs.fs);
+ get_user_ex(vm86regs.gs, &user_vm86->regs.gs);
+
+ get_user_ex(vm86->flags, &user_vm86->flags);
+ get_user_ex(vm86->screen_bitmap, &user_vm86->screen_bitmap);
+ get_user_ex(vm86->cpu_type, &user_vm86->cpu_type);
+ } get_user_catch(err);
+ if (err)
+ return err;
+
+ if (copy_from_user(&vm86->int_revectored,
+ &user_vm86->int_revectored,
+ sizeof(struct revectored_struct)))
+ return -EFAULT;
+ if (copy_from_user(&vm86->int21_revectored,
+ &user_vm86->int21_revectored,
+ sizeof(struct revectored_struct)))
+ return -EFAULT;
+ if (plus) {
+ if (copy_from_user(&vm86->vm86plus, &user_vm86->vm86plus,
+ sizeof(struct vm86plus_info_struct)))
+ return -EFAULT;
+ vm86->vm86plus.is_vm86pus = 1;
+ } else
+ memset(&vm86->vm86plus, 0,
+ sizeof(struct vm86plus_info_struct));
+
+ memcpy(&vm86->regs32, regs, sizeof(struct pt_regs));
+ vm86->user_vm86 = user_vm86;
/*
* The flags register is also special: we cannot trust that the user
* has set it up safely, so this makes sure interrupt etc flags are
* inherited from protected mode.
*/
- VEFLAGS = info->regs.pt.flags;
- info->regs.pt.flags &= SAFE_MASK;
- info->regs.pt.flags |= info->regs32->flags & ~SAFE_MASK;
- info->regs.pt.flags |= X86_VM_MASK;
+ VEFLAGS = vm86regs.pt.flags;
+ vm86regs.pt.flags &= SAFE_MASK;
+ vm86regs.pt.flags |= regs->flags & ~SAFE_MASK;
+ vm86regs.pt.flags |= X86_VM_MASK;
+
+ vm86regs.pt.orig_ax = regs->orig_ax;
- switch (info->cpu_type) {
+ switch (vm86->cpu_type) {
case CPU_286:
- tsk->thread.v86mask = 0;
+ vm86->veflags_mask = 0;
break;
case CPU_386:
- tsk->thread.v86mask = X86_EFLAGS_NT | X86_EFLAGS_IOPL;
+ vm86->veflags_mask = X86_EFLAGS_NT | X86_EFLAGS_IOPL;
break;
case CPU_486:
- tsk->thread.v86mask = X86_EFLAGS_AC | X86_EFLAGS_NT | X86_EFLAGS_IOPL;
+ vm86->veflags_mask = X86_EFLAGS_AC | X86_EFLAGS_NT | X86_EFLAGS_IOPL;
break;
default:
- tsk->thread.v86mask = X86_EFLAGS_ID | X86_EFLAGS_AC | X86_EFLAGS_NT | X86_EFLAGS_IOPL;
+ vm86->veflags_mask = X86_EFLAGS_ID | X86_EFLAGS_AC | X86_EFLAGS_NT | X86_EFLAGS_IOPL;
break;
}
/*
- * Save old state, set default return value (%ax) to 0 (VM86_SIGNAL)
+ * Save old state
*/
- info->regs32->ax = VM86_SIGNAL;
- tsk->thread.saved_sp0 = tsk->thread.sp0;
- tsk->thread.saved_fs = info->regs32->fs;
- tsk->thread.saved_gs = get_user_gs(info->regs32);
+ vm86->saved_sp0 = tsk->thread.sp0;
+ lazy_save_gs(vm86->regs32.gs);
tss = &per_cpu(cpu_tss, get_cpu());
- tsk->thread.sp0 = (unsigned long) &info->VM86_TSS_ESP0;
+ /* make room for real-mode segments */
+ tsk->thread.sp0 += 16;
if (cpu_has_sep)
tsk->thread.sysenter_cs = 0;
load_sp0(tss, &tsk->thread);
put_cpu();
- tsk->thread.screen_bitmap = info->screen_bitmap;
- if (info->flags & VM86_SCREEN_BITMAP)
+ if (vm86->flags & VM86_SCREEN_BITMAP)
mark_screen_rdonly(tsk->mm);
- /*call __audit_syscall_exit since we do not exit via the normal paths */
-#ifdef CONFIG_AUDITSYSCALL
- if (unlikely(current->audit_context))
- __audit_syscall_exit(1, 0);
-#endif
-
- __asm__ __volatile__(
- "movl %0,%%esp\n\t"
- "movl %1,%%ebp\n\t"
-#ifdef CONFIG_X86_32_LAZY_GS
- "mov %2, %%gs\n\t"
-#endif
- "jmp resume_userspace"
- : /* no outputs */
- :"r" (&info->regs), "r" (task_thread_info(tsk)), "r" (0));
- /* we never return here */
-}
-
-static inline void return_to_32bit(struct kernel_vm86_regs *regs16, int retval)
-{
- struct pt_regs *regs32;
-
- regs32 = save_v86_state(regs16);
- regs32->ax = retval;
- __asm__ __volatile__("movl %0,%%esp\n\t"
- "movl %1,%%ebp\n\t"
- "jmp resume_userspace"
- : : "r" (regs32), "r" (current_thread_info()));
+ memcpy((struct kernel_vm86_regs *)regs, &vm86regs, sizeof(vm86regs));
+ force_iret();
+ return regs->ax;
}
static inline void set_IF(struct kernel_vm86_regs *regs)
{
VEFLAGS |= X86_EFLAGS_VIF;
- if (VEFLAGS & X86_EFLAGS_VIP)
- return_to_32bit(regs, VM86_STI);
}
static inline void clear_IF(struct kernel_vm86_regs *regs)
static inline void set_vflags_long(unsigned long flags, struct kernel_vm86_regs *regs)
{
- set_flags(VEFLAGS, flags, current->thread.v86mask);
+ set_flags(VEFLAGS, flags, current->thread.vm86->veflags_mask);
set_flags(regs->pt.flags, flags, SAFE_MASK);
if (flags & X86_EFLAGS_IF)
set_IF(regs);
static inline void set_vflags_short(unsigned short flags, struct kernel_vm86_regs *regs)
{
- set_flags(VFLAGS, flags, current->thread.v86mask);
+ set_flags(VFLAGS, flags, current->thread.vm86->veflags_mask);
set_flags(regs->pt.flags, flags, SAFE_MASK);
if (flags & X86_EFLAGS_IF)
set_IF(regs);
if (VEFLAGS & X86_EFLAGS_VIF)
flags |= X86_EFLAGS_IF;
flags |= X86_EFLAGS_IOPL;
- return flags | (VEFLAGS & current->thread.v86mask);
+ return flags | (VEFLAGS & current->thread.vm86->veflags_mask);
}
static inline int is_revectored(int nr, struct revectored_struct *bitmap)
{
unsigned long __user *intr_ptr;
unsigned long segoffs;
+ struct vm86 *vm86 = current->thread.vm86;
if (regs->pt.cs == BIOSSEG)
goto cannot_handle;
- if (is_revectored(i, &KVM86->int_revectored))
+ if (is_revectored(i, &vm86->int_revectored))
goto cannot_handle;
- if (i == 0x21 && is_revectored(AH(regs), &KVM86->int21_revectored))
+ if (i == 0x21 && is_revectored(AH(regs), &vm86->int21_revectored))
goto cannot_handle;
intr_ptr = (unsigned long __user *) (i << 2);
if (get_user(segoffs, intr_ptr))
return;
cannot_handle:
- return_to_32bit(regs, VM86_INTx + (i << 8));
+ save_v86_state(regs, VM86_INTx + (i << 8));
}
int handle_vm86_trap(struct kernel_vm86_regs *regs, long error_code, int trapno)
{
- if (VMPI.is_vm86pus) {
+ struct vm86 *vm86 = current->thread.vm86;
+
+ if (vm86->vm86plus.is_vm86pus) {
if ((trapno == 3) || (trapno == 1)) {
- KVM86->regs32->ax = VM86_TRAP + (trapno << 8);
- /* setting this flag forces the code in entry_32.S to
- the path where we call save_v86_state() and change
- the stack pointer to KVM86->regs32 */
- set_thread_flag(TIF_NOTIFY_RESUME);
+ save_v86_state(regs, VM86_TRAP + (trapno << 8));
return 0;
}
do_int(regs, trapno, (unsigned char __user *) (regs->pt.ss << 4), SP(regs));
unsigned char __user *ssp;
unsigned short ip, sp, orig_flags;
int data32, pref_done;
+ struct vm86plus_info_struct *vmpi = ¤t->thread.vm86->vm86plus;
#define CHECK_IF_IN_TRAP \
- if (VMPI.vm86dbg_active && VMPI.vm86dbg_TFpendig) \
+ if (vmpi->vm86dbg_active && vmpi->vm86dbg_TFpendig) \
newflags |= X86_EFLAGS_TF
-#define VM86_FAULT_RETURN do { \
- if (VMPI.force_return_for_pic && (VEFLAGS & (X86_EFLAGS_IF | X86_EFLAGS_VIF))) \
- return_to_32bit(regs, VM86_PICRETURN); \
- if (orig_flags & X86_EFLAGS_TF) \
- handle_vm86_trap(regs, 0, 1); \
- return; } while (0)
orig_flags = *(unsigned short *)®s->pt.flags;
SP(regs) -= 2;
}
IP(regs) = ip;
- VM86_FAULT_RETURN;
+ goto vm86_fault_return;
/* popf */
case 0x9d:
else
set_vflags_short(newflags, regs);
- VM86_FAULT_RETURN;
+ goto check_vip;
}
/* int xx */
case 0xcd: {
int intno = popb(csp, ip, simulate_sigsegv);
IP(regs) = ip;
- if (VMPI.vm86dbg_active) {
- if ((1 << (intno & 7)) & VMPI.vm86dbg_intxxtab[intno >> 3])
- return_to_32bit(regs, VM86_INTx + (intno << 8));
+ if (vmpi->vm86dbg_active) {
+ if ((1 << (intno & 7)) & vmpi->vm86dbg_intxxtab[intno >> 3]) {
+ save_v86_state(regs, VM86_INTx + (intno << 8));
+ return;
+ }
}
do_int(regs, intno, ssp, sp);
return;
} else {
set_vflags_short(newflags, regs);
}
- VM86_FAULT_RETURN;
+ goto check_vip;
}
/* cli */
case 0xfa:
IP(regs) = ip;
clear_IF(regs);
- VM86_FAULT_RETURN;
+ goto vm86_fault_return;
/* sti */
/*
case 0xfb:
IP(regs) = ip;
set_IF(regs);
- VM86_FAULT_RETURN;
+ goto check_vip;
default:
- return_to_32bit(regs, VM86_UNKNOWN);
+ save_v86_state(regs, VM86_UNKNOWN);
}
return;
+check_vip:
+ if (VEFLAGS & X86_EFLAGS_VIP) {
+ save_v86_state(regs, VM86_STI);
+ return;
+ }
+
+vm86_fault_return:
+ if (vmpi->force_return_for_pic && (VEFLAGS & (X86_EFLAGS_IF | X86_EFLAGS_VIF))) {
+ save_v86_state(regs, VM86_PICRETURN);
+ return;
+ }
+ if (orig_flags & X86_EFLAGS_TF)
+ handle_vm86_trap(regs, 0, X86_TRAP_DB);
+ return;
+
simulate_sigsegv:
/* FIXME: After a long discussion with Stas we finally
* agreed, that this is wrong. Here we should
* should be a mixture of the two, but how do we
* get the information? [KD]
*/
- return_to_32bit(regs, VM86_UNKNOWN);
+ save_v86_state(regs, VM86_UNKNOWN);
}
/* ---------------- vm86 special IRQ passing stuff ----------------- */
Code Review / kvmfornfv.git / blobdiff